Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

Dataset: 80% C4 (Web Text) and 20% Python Code

Neuron 3400 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1239. Min Act: -0.0000

Data Index: 1139981 (C4 (Web Text))

Max Activating Token Index: 643

Text #1

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1253. Min Act: -0.0001

Data Index: 814973 (C4 (Web Text))

Max Activating Token Index: 509

Text #2

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1297. Min Act: -0.0000

Data Index: 396677 (C4 (Web Text))

Max Activating Token Index: 901

Text #3

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1241. Min Act: -0.0000

Data Index: 428979 (C4 (Web Text))

Max Activating Token Index: 873

Text #4

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1234. Min Act: -0.0001

Data Index: 741274 (C4 (Web Text))

Max Activating Token Index: 699

Text #5

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1254. Min Act: -0.0000

Data Index: 67701 (C4 (Web Text))

Max Activating Token Index: 850

Text #6

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1129. Min Act: -0.0000

Data Index: 432870 (C4 (Web Text))

Max Activating Token Index: 802

Text #7

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1219. Min Act: -0.0000

Data Index: 1304629 (C4 (Web Text))

Max Activating Token Index: 544

Text #8

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1206. Min Act: -0.0000

Data Index: 1377209 (Python Code)

Max Activating Token Index: 155

Text #9

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1108. Min Act: -0.0000

Data Index: 1291220 (C4 (Web Text))

Max Activating Token Index: 838

Text #10

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1080. Min Act: -0.0000

Data Index: 1310745 (C4 (Web Text))

Max Activating Token Index: 136

Text #11

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1139. Min Act: -0.0001

Data Index: 1487250 (Python Code)

Max Activating Token Index: 141

Text #12

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1090. Min Act: -0.0001

Data Index: 1009344 (C4 (Web Text))

Max Activating Token Index: 721

Text #13

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1019. Min Act: -0.0000

Data Index: 687929 (C4 (Web Text))

Max Activating Token Index: 760

Text #14

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1069. Min Act: -0.0001

Data Index: 126117 (C4 (Web Text))

Max Activating Token Index: 584

Text #15

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1025. Min Act: -0.0000

Data Index: 771738 (C4 (Web Text))

Max Activating Token Index: 601

Text #16

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1102. Min Act: -0.0001

Data Index: 434465 (C4 (Web Text))

Max Activating Token Index: 489

Text #17

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1062. Min Act: -0.0000

Data Index: 959947 (C4 (Web Text))

Max Activating Token Index: 486

Text #18

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1039. Min Act: -0.0001

Data Index: 1213887 (C4 (Web Text))

Max Activating Token Index: 592

Text #19

Max Range: 0.1297. Min Range: -0.1297

Max Act: 0.1000. Min Act: -0.0000

Data Index: 429911 (C4 (Web Text))

Max Activating Token Index: 869